34 SCIENTIFIC AMERICAN AUGUST 2005

CREDIT

More complicated than they look,

teeth are actually tiny organs.

If tissue engineers

can manufacture living replacement teeth,

they would blaze a trail for engineering larger organs

while leading dentistry into the age of regenerative medicine

Te s t -Tu b e Te e t h

www.sciam.com SCIENTIFIC AMERICAN 35

CREDIT

By Paul T. Sharpe and Conan S. Young

36 SCIENTIFIC AMERICAN AUGUST 2005

or require major repairs. And then the options are grim: do

without lost teeth or replace them with inert prosthetic ver-

sions. In the Western world, an estimated 85 percent of adults

have had some form of dental treatment. Seven percent have

lost one or more teeth by age 17. After age 50, an average of

12 teeth stand to have been lost.

In theory, a natural tooth made from the patient’s own

tissue and grown in its intended location would make the best

possible replacement, although such bioengineered teeth have

for many years been little more than a dream. Recently, how-

ever, progress in understanding how teeth ﬁ rst develop has

combined with advances in stem cell biology and tissue engi-

neering technology to bring us close to the realization of bio-

logical replacement teeth.

Apart from the potential beneﬁ t to people who need new

teeth, this research also offers two signiﬁ cant advantages for

testing the concept of organ replacement: teeth are easily ac-

cessible, and whereas our quality of life is greatly improved if

we have them, we do not need our teeth to live. These may

seem trivial points, but as the ﬁ rst wave of replacement organs

start to make their way toward the clinic, teeth will serve as

a crucial test of the feasibility of different tissue engineering

techniques. With organs essential to life, doctors will have no

leeway to make mistakes, but mistakes with teeth would not

be life-threatening and could be corrected.

This is not to say that engineering teeth will be simple. Mil-

lions of years of evolution went into establishing the complex

processes that produce organs, teeth included, during embry-

onic development. The challenge for tissue engineers is to rep-

licate those processes, which are tightly controlled by the

growing embryo’s genes. A good way to start learning how to

build teeth, therefore, is to observe how nature does it.

Delicate Dialogue

just six weeks after conception, a human embryo is less

than an inch long and barely beginning to take recognizable

shape. Yet a constant cross talk among its cells is already ini-

tiating and guiding the formation of its teeth. The intricacy

of such signal exchanges is among the reasons that teeth and

other organs cannot as yet be grown entirely in dishes in lab-

oratories. Indeed, scientists may never be able to completely

reproduce these conditions artiﬁ cially. The more we under-

stand these early developmental processes, however, the

greater will be our chances of providing engineered tooth tis-

sues with the most important cues for organ building and

letting nature do the rest.

Most organs, for example, arise through interactions be-

tween two distinct embryonic cell types, epithelial and mes-

enchymal, and teeth are no exception. In the embryo, oral

epithelial cells (which are destined to line oral cavities) send

out the ﬁ rst inductive signals to mesenchymal cells (which will

produce jawbone and soft tissues), instructing them to begin

odontogenesis, or tooth formation. Once the mesenchymal

cells have received their initial instructions, they start sending

signals back to the epithelial cells. This reciprocal exchange

continues throughout embryonic tooth development.

At ﬁ rst, the future tooth is no more than a thickening in

the embryonic oral epithelium. As it grows, the epithelium

begins to penetrate the underlying mesenchymal tissue, which

in turn condenses around the protrusion, forming a tooth bud

by the embryo’s seventh week [see box on opposite page]. As

the epithelium penetrates farther, it wraps itself around the

condensing mesenchyme, eventually forming a bell-shaped

structure, open at its bottom, around 14 weeks. Ultimately,

the epithelium will become the visible outer enamel of the

tooth that erupts from the baby’s gum line some six to twelve

months after birth, and the mesenchymal cells will have

formed the nonvisible parts of the tooth, such as dentin, den-

tal pulp, cementum, and a periodontal ligament that attaches

the tooth to the jawbone.

Even before this tooth begins forming, its shape will be

predetermined by its position. Some of the same epithelial

signals that trigger initiation of odontogenesis also regulate

■ Tissue engineers working toward creating living

replacement teeth take cues from nature as they coax

disparate cell types to form a functional organ.

■ Alternative methods include building teeth from existing

dental cells or growing them from progenitor tissues.

Both approaches have already produced structurally

correct teeth.

■ Remaining challenges include growing roots and

identifying ideal raw materials for bioengineered human

teeth, but progress has been rapid and test-tube teeth

may become the ﬁ rst engineered organs.

Overview/Cutting-Edge Teeth

We take them for granted until they are gone

CARY WOLINSKY (photography); JEN CHRISTIANSEN (photoillustration) (preceding pages)

www.sciam.com SCIENTIFIC AMERICAN 37

HOW NATURE ENGINEERS A TOOTH

TOOTH FORMATION

Just six to seven weeks into human embryonic development, when the whole head is

still taking shape, teeth are also beginning to form. At the location of a future tooth, oral

epithelial tissue thickens slightly and gene activity within its cells causes signals to be sent

to underlying mesenchymal tissue. As the epithelium penetrates farther, mesenchymal

cells respond by emitting their own signals and condensing around the protrusion to form

a tooth bud. By week nine, the epithelium has become a cap atop condensed mesenchyme.

A structure at its center called the enamel knot is now a primary source of signals directing

the activity of both epithelial and mesenchymal cells. At 14 weeks, the tooth germ has a bell

shape comprising differentiating cells called ameloblasts, which will later become enamel, and

odontoblasts, which will form dentin. Roots are the last structures to develop, completing their

formation as the tooth erupts some six to 12 months after birth.

It may look simple from the outside, but on the inside a tooth is

a tiny marvel of design and construction that takes about 14

months to complete in a developing human. Two different types

of primordial embryonic tissue combine to produce a tooth,

and an ongoing molecular dialogue between them directs the

process. Tissue engineers are studying these signals and

steps to understand the cues they need to replicate as they

create living bioengineered replacement teeth.

END RESULT

A living tooth is deﬁ ned as an organ because

it comprises multiple tissue types, each

with an essential function. Enamel, the body’s

hardest mineralized surface, seals and protects

the interior. Dentin, a bony substance, makes

up the bulk of a tooth and serves as a cushion

to resist chewing forces. Pulp, in the center,

contains nourishing blood vessels and nerves

that provide sensory perception. Cementum

forms the hard outer surface of a tooth where it

is not covered by enamel. Periodontal ligament

is a connective tissue that attaches to both the

cementum and the jawbone, anchoring the tooth

in place yet providing some ﬂ exibility.

Embryo at 6 weeks

Thickening: 42–48 days

Tooth bud: 7 weeks

Enamel knot

Bell stage: 14 weeks

Erupted tooth: 6–12 months after birth

Cap stage: 9 weeks

Enamel

Cementum

Gum

Blood vessels

Nerve

Periodontal

ligament

Jawbone

Root

Enamel

Odontoblasts

Oral epithelium

Dentin

Condensing

mesenchyme

Mesenchyme

Signaling

Ameloblasts

Pulp

ANDREW SWIFT

an important category of genes in the jaw mesenchyme.

Known as homeobox genes, they participate in determining

the shape and location of organs and appendages during em-

bryonic development throughout the body. In a developing

human jaw, different homeobox genes are activated in differ-

ent areas, guiding each tooth bud down a pathway to become

a molar, premolar, canine or incisor.

A homeobox gene called Barx1, for example, is switched

on, or expressed, by mesenchymal cells in the positions where

molar teeth will grow. In animal experiments, causing Barx1

to be misexpressed in mesenchyme that would normally form

incisors makes those teeth develop with a molar shape in-

stead. Because the ability to predict and control tooth shape

will be essent ial for the creation of engineered teeth, scientists

can use the activity of genes such as Barx1 as deﬁ nitive predic-

tive markers of future shape when teeth created in the lab are

ﬁ rst growing in culture.

In turn, we must provide the right signals to the developing

teeth at the right time. As early as the 1960s, researchers such

as Shirley Glasstone of Strangeways Research Laboratory in

Cambridge, England, began exploring the possibility of grow-

ing teeth by experimenting with mouse tissues. In seminal

studies performed over the next three decades, tiny pieces of

embryonic mouse dental epithelium and dental mesenchyme

were brought together and then either grown in a tissue culture

dish or surgically implanted in the body of a host where the

recombined tissues would receive a blood supply. These ex-

periments demonstrated that such embryonic tooth primordia

could continue to develop as if they were still in the embryo,

producing dentin and enamel. Their development arrests ear-

ly, however, and they do not ultimately yield fully formed

teeth. Something is missing from their environment.

The growth factors and other signals required to complete

tooth formation in an embryo most likely come from sur-

rounding jaw tissue. Thus, transplanting tooth primordia into

the jaw to ﬁ nish developing would seem to be a simple solu-

tion. When replacement teeth are engineered, for instance,

they will ideally be grown in their permanent location so that

they can create nerve and blood vessel connections and phys-

ically attach themselves to the jawbone. The adult jaw is a

vastly different environment from the embryonic version,

however, and scientists have been unsure whether it would

provide the correct signals to a developing tooth.

Moreover, tooth primordia must be constructed from the

right combination of cells to reproduce natural tooth mate-

rial and structure. Being able to use cells from a patient’s own

body would be preferable to using embryonic cells because

the patient’s own tissue would not be perceived as foreign and

so would not provoke an immune response.

Three key milestones must therefore be reached to estab-

lish whether engineering replacement biological teeth is pos-

sible. Sources of cells that can form teeth and are easily ob-

tained from patients themselves must be identiﬁ ed. The teeth

produced from these cells must be able to develop in the envi-

ronment of the adult jaw, producing roots that are attached to

Cells Reunite to Form Teeth

Tooth cells taken from adolescent pigs and seeded onto a

biodegradable scaffold are visible in blue along its edges

after one week of incubation (top left). Following 25

weeks of growth (top right), the scaffold has dissolved and new

dental pulp, enamel, and dentin have taken its place. In a series

of such experiments, tiny toothlike structures grew amid the

new tissues. Correct tooth-tissue organization (bottom left),

including a pre-root structure known as Hertwig’s epithelial

root sheath (Hers), was observed in 15 to 20 percent of the

miniature teeth. In other instances, the tooth structure was

incorrect or incomplete (bottom right). These bioengineered

teeth nonetheless seem to conﬁ rm that disaggregated dental

cells can reorganize themselves into larger dental tissues.

Pulp

Scaffold

Hers

Enamel

Dentin

Predentin

Odontoblasts

Dentin

Tooth cells

Pulp

Enamel

REPRINTED FROM DEVELOPMENTAL ANALYSIS AND COMPUTER MODELLING OF BIOENGINEERED TEETH, BY C. S. YOUNG ET AL. IN ARCHIVES OF ORAL BIOLOGY,

VOL. 50, PAGES 259–265; 2005. WITH PERMISSION FROM ELSEVIER (top left and right; bottom right); CONAN S. YOUNG (bottom left)

38 SCIENTIFIC AMERICAN AUGUST 2005

www.sciam.com SCIENTIFIC AMERICAN 39

the bone by a functional periodontal ligament. And the shape

and size of these biological teeth must be predictable and con-

trollable so that they can be made to match the patient’s own

teeth. These are ambitious goals, but considerable progress

toward each is being made by different research groups using

somewhat disparate approaches.

Building Bioteeth

in the late 1980s organ transplant surgeon Joseph P.

Vacanti of Harvard Medical School and polymer chemist Rob-

ert S. Langer of the Massachusetts Institute of Technology

conceived the idea of placing the cells of an organ or tissue on

a prefabricated biodegradable scaffold with the goal of gener-

ating tissues and organs for transplantation [see “Artiﬁ cial

Organs,” by Robert S. Langer and Joseph P. Vacanti; Scien-

tiﬁ c American, September 1995]. In simpliﬁ ed terms, their

approach was based on the fact that living tissues are made of

cells constantly signaling to one another and often moving

around within a three-dimensional community of sorts. Each

cell seems to know its place and role in the larger collective that

forms and maintains a functional tissue. Therefore, if the right

mix of dissociated cells is reaggregated within a scaffold that

replicates their natural 3-D environment, the cells should in-

stinctively reform the tissue or organ to which they belong.

Vacanti and Langer’s early successes regenerating pieces of

liver tissue from liver cells using this scaffold-based strategy

have since led to widespread experimentation with the tech-

nique to produce other complex tissues, such as heart muscle,

intestine, mineralized bone and now teeth. Pamela C. Yelick

and John D. Bartlett of the Forsyth Institute in Boston began

working with Vacanti in 2000 to investigate the feasibility of

engineering teeth this way by focusing on pigs, which, like

humans, produce two sets of teeth over their lifetime.

One of us (Young) also took part in these experiments for

which raw material was derived from the unerupted third mo-

lars (“wisdom teeth”) of six-month-old pigs. To obtain a het-

erogeneous random mixture of dental enamel epithelial and

pulp mesenchymal cells, the pig teeth were broken into tiny

pieces and then further dissolved using enzymes. Tooth-

shaped scaffolds were made from biodegradable polyester

plastics and coated with a substance that makes the plastic

sticky so cells can adhere to it. The cell mixtures were seeded

into the scaffolds, and the constructs were surgically implant-

ed into rat hosts, wrapped in omentum, a fatty white material

rich in blood vessels that surrounds the intestines. This step is

important because the developing tooth tissues require an

ample blood supply to provide them with nutrients and oxygen

while they grow.

Initially the scaffolds provided support for the cells, but

later they dissolved as intended and were replaced by new tis-

sue. When the implants were examined after 20 to 30 weeks,

tiny toothlike structures were visible within the conﬁ nes of the

original scaffold. Their shape and the organization of their

tissues resembled the crowns of natural teeth [see box on op-

posite page]. They also included most of the tissues that make

up a normal tooth, demonstrating for the ﬁ rst time that enam-

el, dentin, pulp, and features that appeared to be developing

tooth roots could be regenerated on scaffolds.

It seemed that mixtures of dental cells could reorganize

themselves on scaffolds into arrangements that favor forma-

tion of mineralized enamel, dentin and soft tooth tissue. An-

other possible explanation for these exciting results, of course,

was that the random arrangement of cells seeded onto the scaf-

fold favored tooth tissue development only by chance. The

Forsyth group therefore tested these possibilities in a new

study using dental epithelial and mesenchymal cells isolated

from the ﬁ rst, second and third molars of rats. This time, how-

ever, the cells were grown and their numbers expanded in tis-

sue culture for six days before their being seeded onto scaffolds

and implanted in rat hosts. After 12 weeks’ growth, the result-

ing tissues were extracted and examined. Once again, small

tooth structures consisting of enamel, dentin and pulp tissue

were observed to have formed within the original scaffold.

These new results were encouraging because they lent some

weight to the previous evidence that cells can reorganize them-

Each cell seems to know

its place in

the larger collective.

PAUL T. SHARPE and CONAN S. YOUNG met two years ago at a

tooth and bone conference where they discovered a shared

fondness for mountain biking and soccer (one calls it “foot-

ball”), despite their differing approaches to bioengineering

teeth. Sharpe established and heads the department of cranio-

facial development at Guy’s Hospital in London and is also Dick-

inson Professor of Craniofacial Biology at King’s College Lon-

don. In 2002 he founded Odontis Ltd., a biotechnology company

devoted to growing human teeth and bone by emulating their

formative processes in a developing embryo. Young is an in-

structor in oral and developmental biology at the Harvard

School of Dental Medicine and a staff scientist at the Forsyth

Institute in Boston, where he is working toward growing teeth

from cells seeded onto biodegradable scaffolds.

THE AUTHORS

40 SCIENTIFIC AMERICAN AUGUST 2005

selves into tooth-forming conﬁ gurations. Moreover, the cells

did not appear to have been adversely affected by being ex-

panded in culture—a process that will be essential in engineer-

ing human replacement teeth because tissue engineers would

probably have to craft a replacement tooth from small samples

of the patient’s own cells. And, ﬁ nally, the experiment demon-

strated that tooth regeneration is possible in a second mam-

mal, making the success of a similar approach in humans more

likely.

Although the Forsyth team was able to generate most of

the desired tissue types with cells from an adult source, those

tissues organized themselves into the proper arrangement for

a natural tooth only 15 to 20 percent of the time. The group is

therefore continuing to work on methods of more precisely

placing different dental cell types within scaffolds to achieve

a more accurate tooth structure.

At the same time, the team is exploring the possibility that

the new tooth tissues observed in these experiments might not

have been produced solely by reorganization of the dissociated

dental cells. Instead the third molar tooth buds that provided

cells to seed the scaffolds might have contained hidden stem

cells—potent progenitors of other cell types—that were re-

sponsible for forming the new tissue. If true, this would mean

that new dental stem cells capable of producing nearly all the

dental tissue types required for bioengineering teeth might

exist within teeth themselves, at least until early adulthood,

when wisdom teeth erupt. Such versatile adult dental stem cells

would certainly speed efforts to generate teeth on scaffolds,

and they might also facilitate the tooth-engineering approach

used by the Sharpe group at King’s College London.

Teeth from Scratch

rather than attempting to build adult teeth from

their const it uent cells, one of us (Sharpe) is pursuing a strategy

based more closely on reproducing the natural processes of

embryonic tooth development described earlier. In essence, the

method requires an u nderstanding of the basic principles con-

trolling early tooth formation and a source of cells to play the

roles of embryonic oral epithelium and mesenchyme.

To date, the Sharpe group has experimented primarily

with mouse cells, using both stem cells and ordinary cells,

from embryonic as well as adult sources, to test the potential

of various cell types to produce replacement teeth. In most

cases, the group began by aggregating mesenchymal cells in a

centrifuge until they formed a small solid mass. This pellet was

then covered in epithelium and cultured for several days, while

the gene activity in its tissues was monitored for indications of

early tooth development. Next, these tooth primordia were

implanted into the bodies of animal hosts in locations where

they could receive a nourishing blood supply, such as the kid-

ney of a mouse, and left to grow for about 26 days.

In the course of these experiments, clear tooth formation

was observed but only when the epithelium came from an em-

bryonic source and the mesenchymal cell populations con-

tained at least some stem cells. When stem cells from adult

bone marrow took the place of oral mesenchyme, for example,

the transplanted constructs produced structurally correct

teeth. Thus, it seems embryonic mesenchyme can be replaced

with adult stem cells to generate new teeth.

Unfortunately, many years of experiments have established

that embryonic epithelium contains a unique set of signals for

odontogenesis that disappear from the mouth after birth. The

Sharpe group is continuing to seek an effective population of

substitute cells that could be derived from an adult source.

Still, the results achieved with primordia made from the com-

bination of adult stem cells and embryonic oral epithelium

have been extremely encouraging.

Signiﬁ cantly, these teeth were also in the normal size range

for mouse teeth, they were surrounded by new bone and con-

nective tissue, and they showed the earliest signs of root forma-

tion. The next step was to see whether such explants could also

form teeth in the mouth. In the embryonic jaw, soft tissues,

teeth and bone are all developing together without external

stresses such as chewing and talking, whereas the adult jaw is

a hard, busy place. No one knew whether it would provide the

necessary signals for teeth to form and integrate themselves

into the environment as they would in an embryo.

To ﬁ nd out, the Sharpe group extracted tooth buds from

embryonic mice, then transplanted them into the mouths of

adult mice. Small incisions were made in the soft tissue of the

upper jaw of the host mice, in a region known as the diastema

between the molars and incisors where normally there are no

teeth. The embryonic tooth primordia were inserted into

these pockets and sealed in place with surgical glue. After-

ward, the mice were fed a soft diet and the transplants moni-

tored. Just three weeks later teeth could be clearly identiﬁ ed

in the diastema. They had formed in the correct orientation,

were of appropriate size for the mice, and were attached to

underlying bone by soft connective tissue [see illustration on

opposite page].

No one knew whether the

adult jaw would provide signals for

teeth to form.

www.sciam.com SCIENTIFIC AMERICAN 41

Remarkably, it appears that the adult mouth can provide a

suitable environment for tooth development. That is just one

of the three milestones toward engineering replacement teeth

that we identiﬁ ed earlier, however. The road to human bioen-

gineered teeth may yet have a few twists.

On the Cusp

compared with efforts to engineer other organs,

teeth have made considerable progress in a short time. The

overall challenge remains developing methods that are simple

yet controlled.

Another of the targets that we established, the ability to

predict and control tooth size and shape, is close. In cultured

primordia, molar and incisor tooth germs can easily be dis-

tinguished by their appearance and their gene activity, al-

though other shapes found in the human mouth, such as pre-

molars and canines, are more difﬁ cult.

The teeth grown from embryonic primordia in the mouths

of adult mice by the Sharpe group displayed shapes appropri-

ate to their original locations in the embryo—molar primor-

dia grew into molar-shaped teeth, for example. Because shape

signals are received at the very start of natural tooth develop-

ment, the embryonic tooth germs were already programmed.

Tissue engineers need to better understand these initial shape

signals to induce them in human bioteeth.

To date, the teeth generated by any of the tissue engineer-

ing methods we have described have not developed roots. In

truth, both root development and the stimuli that initiate tooth

eruption are complex and still little understood. Roots are the

last part of teeth to form, completing their development during

the eruption process, and more research is needed to under-

stand what conditions would best favor their creation in re-

placement teeth. Another unknown is how long engineered

human teeth would take to fully form in an adult mouth. Hu-

mans’ second set of “adult” teeth also begins developing in the

embryo, yet those teeth take six to seven years to ﬁ nally

erupt—or 20 years in the case of wisdom teeth. Our experience

with tooth generation in animals suggests that an engineered

human tooth would form far more quickly, but we do not

know if it might take longer to fully mature and its enamel to

completely harden.

Of course, most research into bioengineered tooth pro-

duction is also working toward ﬁ nding an effective and easily

accessible source of the patient’s own cells to use as raw mate-

rial. Immune rejection would be avoided, and because tooth

size, shape and color are genetically determined, the engi-

neered teeth would more closely match the patient’s natural

teeth. The Sharpe group has found that adult mesenchymal

stem cells derived from bone marrow (but also possibly ob-

tainable from fat) can replace embryonic mesenchyme in the

tooth formation process. A substitute for embryonic epithe-

lium has yet to be identiﬁ ed, although purported adult stem

cells have been discovered in other tissues with epithelial ori-

gins, such as skin and hair. These or some other adult cell type

may prove effective, perhaps with the aid of gene manipulation

to induce the appropriate initiating signals for odontogenesis.

Of the several potential cell sources, teeth themselves may

be the most convenient. The Forsyth group’s results suggest

that stem cells capable of forming tooth tissues, including

enamel, could be present within teeth. Researchers elsewhere

have also shown that dentin and other tooth tissues experience

some natural regeneration after injury, which, too, suggests

the presence of progenitor cells capable of generating a variety

of tooth tissues. Thus, the possibility exists of someday soon

fashioning new teeth from old.

MORE TO EXPLORE

Tissue Engineering: The Challenges Ahead. Robert S. Langer and

Joseph P. Vacanti in Scientiﬁ c American, Vol. 280, No. 4, pages 86–89;

April 1999.

Tissue Engineering of Complex Tooth Structures on Biodegradable

Polymer Scaffolds. Conan S. Young, Shinichi Terada, Joseph P. Vacanti,

Masaki Honda, John D. Bartlett and Pamela C. Yelick in Journal of Dental

Research, Vol. 81, No. 10, pages 695–700; October 2002.

Bioengineered Teeth from Cultured Rat Tooth Bud Cells. Monica T.

Duailibi, Silvio E. Duailibi, Conan S. Young, John D. Bartlett, Joseph P.

Vacanti and Pamela C. Yelick in Journal of Dental Research, Vol. 83,

No. 7, pages 523–528; July 2004.

Stem Cell Based Tissue Engineering of Murine Teeth. A. Ohazama,

S.A.C. Modino, I. Miletich and P. T. Sharpe in Journal of Dental Research,

Vol. 83, No. 7, pages 518–522; July 2004.

The Cutting Edge of Mammalian Development: How the Embryo

Makes Teeth. Abigail S. Tucker and Paul T. Sharpe in Nature Reviews

Genetics, Vol. 5, No. 7, pages 499–508; July 2004.

MOUSE TOOTH grown from transplanted molar primordia

in the upper jaw of a host mouse demonstrates that new

teeth can develop in the adult mouth. The tooth at center in this cross

section of the jaw’s diastema region has broken through the gum line

(a second tooth above it and to the right is still forming). Pulp is visible

inside the emerged tooth. Red stain colors dental hard tissues,

highlighting enamel and dentin. Although lacking roots, the tooth is

attached to surrounding jawbone by soft connective tissue.

Connective

tissue

Jawbone

Dentin Pulp

Enamel

Gum

Tooth

MATT BRADMAN